Hard Drive Issues After Scheduled Parity Check

[About7Deaths]

After the parity check started itself, it failed to complete after a few hours and my main drive (the other drive being the parity drive) was taken offline and SMART report was unavailable. Doing some reading, I downloaded the current diagnostics zip, disabled auto start of VMs and auto start of array, and then shut down the system. I rebooted but no cigar. I shut down the server and disconnected the wires, I rebooted and logged into the portal on my laptop. I then reconnected the one problematic drive see see if it would pop up, after 5+ minutes it repopulated the array and I was able to run a smart test (although it may not have completed all the way, it was hard to tell). The only notable thing I noticed is that Current_Pending_Sector had a RAW_VALUE of 1. Attached is the SMART report. After reading the report everything seemed fine (to my inexperienced eye), however, upon reattaching everything and rebooting I am still unable to see or attach the problem drive.

han1-smart-20210203-1134.zip

[JorgeB]

14 minutes ago, About7Deaths said:

upon reattaching everything and rebooting I am still unable to see or attach the problem drive.

Do you mean the drive is not available in Unraid? This can only be a drive or connection problem, if it ends showing up you should also run an extended SMART test to confirm if there are pending sectors or not.

[About7Deaths]

3 minutes ago, JorgeB said:

Do you mean the drive is not available in Unraid? This can only be a drive or connection problem, if it ends showing up you should also run an extended SMART test to confirm if there are pending sectors or not.

Correct, the drive is not available. The SMART test I provided was what was available to download after selecting the extended test option from the portal. However, it is possible that the test failed to run in its entirety.

[About7Deaths]

The parity drive seems to be fine and I have backup spare drives. Is it to the point where I should ditch the drive and rebuild the array with the spare? (The current two drives are the same age at 5+ years.) The rebuild would be around 3 TB. I actually have two extra drives at 3 TB; would it be advisable to plug them both in (making it 3 total at 3/3/3 TB) and rebuild instead of just doing 1 to 1?

Edited February 3, 2021 by About7Deaths

[JorgeB]

12 hours ago, About7Deaths said:

after selecting the extended test option

The test didn't finish.

 

12 hours ago, About7Deaths said:

The parity drive seems to be fine and I have backup spare drives. Is it to the point where I should ditch the drive and rebuild the array with the spare?

Yes, if the original drive is not detected or the SMART test fails.

 

12 hours ago, About7Deaths said:

The rebuild would be around 3 TB.

Rebuild would be the size of the new drive.

 

12 hours ago, About7Deaths said:

would it be advisable to plug them both in (making it 3 total at 3/3/3 TB) and rebuild instead of just doing 1 to 1?

Don't understand what you mean.

[About7Deaths]

On 2/4/2021 at 1:47 AM, JorgeB said:

Don't understand what you mean.

If I were to plug in two drives instead of one, would it split the data between the two drives when rebuilding? Whereas if I only put in a single drive for rebuild it would definitely fill the whole drive.

[About7Deaths]

On 2/4/2021 at 1:47 AM, JorgeB said:

Rebuild would be the size of the new drive.

For this I meant how long does a small 3TB rebuild usually take time-wise.

[About7Deaths]

I went ahead and plugged in the one drive and am currently pre-clearing it. Although, thinking back on this I'm surprised that I have the settings to do a parity check once a year and it coincidentally failed the one day that I have it automatically check? Seems odd. It had checked a month prior because of a power failure and was fine.

[JonathanM]

16 minutes ago, About7Deaths said:

I have the settings to do a parity check once a year

That's probably a little long of an interval. The consensus is once a month. A parity check with zero errors is the way to know with some certainty that a drive failure can be rebuilt accurately.

 

If you aren't going to do parity checks more often, at least do long smart tests on all your drives, that way at least drive failures have a chance of being caught before multiples sneak up on you and preclude the possibility of rebuilding successfully.

[About7Deaths]

1 minute ago, jonathanm said:

That's probably a little long of an interval. The consensus is once a month. A parity check with zero errors is the way to know with some certainty that a drive failure can be rebuilt accurately.

 

If you aren't going to do parity checks more often, at least do long smart tests on all your drives, that way at least drive failures have a chance of being caught before multiples sneak up on you and preclude the possibility of rebuilding successfully.

 

Definitely not best practice on my part but I usually did it manually from time to time. I just set it for a year in case I managed to totally forget (which I never did). I just find it odd how after the parity check I am not even able to detect the drive after plugging it in, let alone running tests on it.

[JorgeB]

10 hours ago, About7Deaths said:

If I were to plug in two drives instead of one, would it split the data between the two drives when rebuilding?

No, you can only rebuild a disable drive to a single device, same size or larger than the original, and can't be larger than parity.

 

3TB rebuild should take around 5 or 6 hours, depending on the disks and other the hardware. 

[About7Deaths]

I rebuilt the array and all works as expected; thank you all for the advice.